1 line
60 KiB
JSON
1 line
60 KiB
JSON
{"repo": "assert-kth/flames", "files": {"raw-validation-results/sb-heists/inference-single-turn-deepseek.ipynb": {"8dadcf88b53331dd1dd1d2be4dfb54f21ebf7a95": {"t": 1785745512.0114388, "n": 17721, "h": "fcb5e11e0a1ad825", "c": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"86039af7\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"\\n\",\n \"# Requires: pip install openai python-dotenv\\n\",\n \"import os, time, re, json\\n\",\n \"from typing import Optional\\n\",\n \"from dotenv import load_dotenv\\n\",\n \"\\n\",\n \"load_dotenv()\\n\",\n \"\\n\",\n \"try:\\n\",\n \" from openai import OpenAI\\n\",\n \"except Exception as e:\\n\",\n \" raise RuntimeError(\\\"Install the OpenAI SDK first: pip install openai\\\") from e\\n\",\n \"\\n\",\n \"DEEPSEEK_API_KEY = os.environ.get(\\\"DEEPSEEK_API_KEY\\\", \\\"\\\")\\n\",\n \"DEEPSEEK_BASE_URL = os.environ.get(\\\"DEEPSEEK_BASE_URL\\\", \\\"https://api.deepseek.com\\\")\\n\",\n \"DEEPSEEK_MODEL = os.environ.get(\\\"DEEPSEEK_MODEL\\\", \\\"deepseek-chat\\\") # or your preferred\\n\",\n \"\\n\",\n \"if not DEEPSEEK_API_KEY:\\n\",\n \" raise RuntimeError(\\\"Please set DEEPSEEK_API_KEY in your environment (or .env).\\\")\\n\",\n \"\\n\",\n \"client = OpenAI(api_key=DEEPSEEK_API_KEY, base_url=DEEPSEEK_BASE_URL)\\n\",\n \"\\n\",\n \"def _chat_once(prefix: str, suffix: str, temperature=0.8, top_p=0.95, max_tokens=128) -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Sends a single chat completion asking for ONLY the Solidity predicate (no require(...);).\\n\",\n \" Uses a FIM-like instruction by giving prefix/suffix and a <FILL_ME> placeholder.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" prompt = (\\n\",\n \" \\\"You are filling a single-line Solidity predicate to go inside require(...).\\\\n\\\"\\n\",\n \" \\\"Given a source split into PREFIX and SUFFIX around a placeholder `require(<FILL_ME>);`, \\\"\\n\",\n \" \\\"produce ONLY the predicate expression for the require (no semicolon, no backticks, no commentary).\\\\n\\\"\\n\",\n \" \\\"Examples:\\\\n\\\"\\n\",\n \" \\\" - Output like: msg.sender == owner\\\\n\\\"\\n\",\n \" \\\" - NOT like: require(msg.sender == owner);\\\\n\\\"\\n\",\n \" \\\" - NOT like: ```solidity ... ```\\\\n\\\"\\n\",\n \" \\\"\\\\n\\\"\\n\",\n \" \\\"PREFIX:\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" f\\\"{prefix}\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" \\\"SUFFIX:\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" f\\\"{suffix}\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" )\\n\",\n \" resp = client.chat.completions.create(\\n\",\n \" model=DEEPSEEK_MODEL,\\n\",\n \" messages=[\\n\",\n \" {\\\"role\\\": \\\"system\\\", \\\"content\\\": \\\"Return only the Solidity predicate expression.\\\"},\\n\",\n \" {\\\"role\\\": \\\"user\\\", \\\"content\\\": prompt},\\n\",\n \" ],\\n\",\n \" temperature=temperature,\\n\",\n \" top_p=top_p,\\n\",\n \" max_tokens=max_tokens,\\n\",\n \" stop=None,\\n\",\n \" )\\n\",\n \" return (resp.choices[0].message.content or \\\"\\\").strip()\\n\",\n \"\\n\",\n \"def _extract_predicate(text: str) -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Normalize model output to a single-line predicate:\\n\",\n \" - Strip code fences/backticks/semicolons\\n\",\n \" - If it accidentally returned require(...), extract the (...)\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" t = text.strip()\\n\",\n \" t = re.sub(r\\\"^```[a-zA-Z]*\\\\s*\\\", \\\"\\\", t).strip()\\n\",\n \" t = re.sub(r\\\"```$\\\", \\\"\\\", t).strip()\\n\",\n \" # require(...) wrapper?\\n\",\n \" m = re.search(r\\\"require\\\\s*\\\\(\\\\s*(.+?)\\\\s*\\\\)\\\\s*;?\\\\s*$\\\", t, flags=re.S)\\n\",\n \" if m:\\n\",\n \" return m.group(1).strip()\\n\",\n \" # strip trailing semicolon if present\\n\",\n \" t = re.sub(r\\\";+$\\\", \\\"\\\", t).strip()\\n\",\n \" # collapse lines\\n\",\n \" t = \\\" \\\".join(t.splitlines()).strip()\\n\",\n \" return t\\n\",\n \"\\n\",\n \"def deepseek_generate_fim(prefix: str, suffix: str) -> str:\\n\",\n \" # lightweight retry/backoff\\n\",\n \" delays = [0, 1, 2, 4]\\n\",\n \" for d in delays:\\n\",\n \" if d: time.sleep(d)\\n\",\n \" try:\\n\",\n \" raw = _chat_once(prefix, suffix)\\n\",\n \" return _extract_predicate(raw)\\n\",\n \" except Exception:\\n\",\n \" if d == delays[-1]:\\n\",\n \" raise\\n\",\n \" continue\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"8be587ac\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Injection utilities: build masked prefix/suffix at a location (single-turn)\\n\",\n \"import validation_library as veri\\n\",\n \"\\n\",\n \"def _insert_mask_at_line(src: str, line_no_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert `require(<FILL_ME>);` BEFORE 1-based line_no_1based.\\n\",\n \" Returns (prefix, suffix, indent_str).\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" lines = src.splitlines()\\n\",\n \" idx = max(0, min(len(lines), line_no_1based - 1))\\n\",\n \" # Determine indentation: use target line's indentation if exists\\n\",\n \" if 0 <= idx < len(lines):\\n\",\n \" m = re.match(r\\\"^(\\\\s*)\\\", lines[idx])\\n\",\n \" indent = m.group(1) if m else indent_fallback\\n\",\n \" else:\\n\",\n \" indent = indent_fallback\\n\",\n \" masked = f\\\"{indent}require(<FILL_ME>);\\\"\\n\",\n \" prefix = \\\"\\\\n\\\".join(lines[:idx] + [masked])\\n\",\n \" suffix = \\\"\\\\n\\\".join(lines[idx:])\\n\",\n \" return prefix, suffix, indent\\n\",\n \"\\n\",\n \"def _insert_mask_at_function_entry(src: str, vl_line_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert mask as the first statement of the function that contains vl_line_1based.\\n\",\n \" Uses veri.find_function_bounds to resolve the function block.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" pre_line_0based, post_line_0based = veri.find_function_bounds(src, vl_line_1based) # returns 0-based\\n\",\n \" # we want to insert just after the opening brace -> typical pre = opening brace line\\n\",\n \" insert_at = pre_line_0based + 1 # 1-based\\n\",\n \" return _insert_mask_at_line(src, insert_at, indent_fallback=indent_fallback)\\n\",\n \"\\n\",\n \"def _insert_mask_at_function_exit(src: str, vl_line_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert mask just before the closing brace of the function that contains vl_line_1based.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" pre_line_0based, post_line_0based = veri.find_function_bounds(src, vl_line_1based)\\n\",\n \" insert_at = post_line_0based + 1 # convert to 1-based\\n\",\n \" return _insert_mask_at_line(src, insert_at, indent_fallback=indent_fallback)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"b773c049\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Core: single-turn synthesizer for {V, alpha, omega}\\n\",\n \"def synthesize_require_statement(\\n\",\n \" contract_text: str,\\n\",\n \" line_no: int, # VL reference line (1-based)\\n\",\n \" injection: str = \\\"V\\\", # \\\"V\\\" (target line), \\\"alpha\\\" (function entry), \\\"omega\\\" (function exit)\\n\",\n \" indent_fallback: str = \\\" \\\",\\n\",\n \" extra_hint: Optional[str] = None,\\n\",\n \") -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Returns ONLY the predicate (no semicolon). We will wrap as require({pred});\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" if injection == \\\"V\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_line(contract_text, line_no, indent_fallback)\\n\",\n \" elif injection == \\\"alpha\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_function_entry(contract_text, line_no, indent_fallback)\\n\",\n \" elif injection == \\\"omega\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_function_exit(contract_text, line_no, indent_fallback)\\n\",\n \" else:\\n\",\n \" raise ValueError(\\\"injection must be one of: 'V', 'alpha', 'omega'\\\")\\n\",\n \"\\n\",\n \" if extra_hint:\\n\",\n \" prefix = f\\\"// Hint: {extra_hint}\\\\n\\\" + prefix\\n\",\n \"\\n\",\n \" predicate = deepseek_generate_fim(prefix, suffix)\\n\",\n \" return predicate\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"19c25c07\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Seed loader (auto from your aggregated VL artifact)\\n\",\n \"# This replaces the previous seed_contracts-based cell.\\n\",\n \"\\n\",\n \"import os, json\\n\",\n \"from collections import defaultdict\\n\",\n \"\\n\",\n \"# We’ll auto-pick one of these paths unless SEED_VL_JSON is set.\\n\",\n \"CANDIDATES = [\\n\",\n \" os.environ.get(\\\"SEED_VL_JSON\\\", \\\"\\\").strip(),\\n\",\n \" \\\"reports/aggregated/DeepSeek/contract_VL.json\\\",\\n\",\n \" \\\"reports/aggregated/DeepSeek/contracts_VL.json\\\",\\n\",\n \" \\\"reports/aggregated/contract_VL.json\\\",\\n\",\n \"]\\n\",\n \"\\n\",\n \"seed_path = next((p for p in CANDIDATES if p and os.path.exists(p)), None)\\n\",\n \"if not seed_path:\\n\",\n \" raise RuntimeError(\\n\",\n \" \\\"Could not find a VL seed JSON. \\\"\\n\",\n \" \\\"Set SEED_VL_JSON to your file or generate reports/aggregated/DeepSeek/contract_VL.json first.\\\"\\n\",\n \" )\\n\",\n \"\\n\",\n \"with open(seed_path, \\\"r\\\", encoding=\\\"utf-8\\\") as f:\\n\",\n \" raw = json.load(f)\\n\",\n \"\\n\",\n \"# raw is expected to be: { name: [ [contract_text, { 'VL':[line, guard], 'pre':[line, ''], 'post':[line, ''] } ], ... ] }\\n\",\n \"# (Tuples become lists in JSON; guards may be empty strings.)\\n\",\n \"contracts_with_results = defaultdict(list)\\n\",\n \"\\n\",\n \"def _read_line_pair(x):\\n\",\n \" # Accept [line, guard] or just line\\n\",\n \" if isinstance(x, list) and x:\\n\",\n \" return int(x[0]), (x[1] if len(x) > 1 else \\\"\\\")\\n\",\n \" if isinstance(x, int):\\n\",\n \" return int(x), \\\"\\\"\\n\",\n \" return None, \\\"\\\"\\n\",\n \"\\n\",\n \"loaded = 0\\n\",\n \"for name, entries in raw.items():\\n\",\n \" for entry in entries:\\n\",\n \" # entry may be [contract_text, annotations] or a dict with same fields\\n\",\n \" if isinstance(entry, list) and len(entry) == 2:\\n\",\n \" contract_text, ann = entry\\n\",\n \" elif isinstance(entry, dict):\\n\",\n \" contract_text = entry.get(\\\"contract\\\") or entry.get(\\\"contract_text\\\") or entry.get(\\\"0\\\")\\n\",\n \" ann = entry.get(\\\"annotations\\\") or entry.get(\\\"1\\\") or {k: entry.get(k) for k in (\\\"VL\\\",\\\"pre\\\",\\\"post\\\")}\\n\",\n \" else:\\n\",\n \" continue\\n\",\n \"\\n\",\n \" if not isinstance(contract_text, str) or not contract_text.strip():\\n\",\n \" continue\\n\",\n \"\\n\",\n \" vl_line, vl_guard = _read_line_pair((ann or {}).get(\\\"VL\\\"))\\n\",\n \" pre_line, pre_guard = _read_line_pair((ann or {}).get(\\\"pre\\\"))\\n\",\n \" post_line, post_guard = _read_line_pair((ann or {}).get(\\\"post\\\"))\\n\",\n \"\\n\",\n \" # Fallback: recompute pre/post lines if missing using your helper\\n\",\n \" if not isinstance(vl_line, int) or vl_line <= 0:\\n\",\n \" continue\\n\",\n \" try:\\n\",\n \" pre0, post0 = veri.find_function_bounds(contract_text, vl_line) # 0-based\\n\",\n \" if not isinstance(pre_line, int) or pre_line <= 0:\\n\",\n \" pre_line = pre0 + 1 # store function-start (brace) line (0-based+1)\\n\",\n \" # We actually insert at first statement; later logic uses VL+injection to compute exact spot.\\n\",\n \" if not isinstance(post_line, int) or post_line <= 0:\\n\",\n \" post_line = post0 + 1 # closing brace line (0-based+1)\\n\",\n \" except Exception:\\n\",\n \" continue\\n\",\n \"\\n\",\n \" contracts_with_results[name].append((\\n\",\n \" contract_text,\\n\",\n \" {\\n\",\n \" \\\"VL\\\": (vl_line, vl_guard or \\\"\\\"),\\n\",\n \" \\\"pre\\\": (pre_line, pre_guard or \\\"\\\"),\\n\",\n \" \\\"post\\\":(post_line, post_guard or \\\"\\\"),\\n\",\n \" }\\n\",\n \" ))\\n\",\n \" loaded += 1\\n\",\n \"\\n\",\n \"print(f\\\"[seed] Loaded {loaded} contract entries from {seed_path}\\\")\\n\",\n \"len(contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"a8d85734\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Output directory for SINGLE-TURN results\\n\",\n \"os.makedirs(\\\"reports/isolated/DeepSeek\\\", exist_ok=True)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"8461a3bb\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% SINGLE-TURN: Vulnerable-line (V) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_V = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"VL single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" # SINGLE-TURN: use ORIGINAL contract (no prior injections)\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"V\\\")\\n\",\n \" results_V.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, vl_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"VL\\\"]\\n\",\n \"results_V += [''] * (len(all_contracts) - len(results_V))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, vl_line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_V, mapping), flat_entries):\\n\",\n \" annotations[\\\"VL\\\"] = (vl_line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"# Minimal VL-only drop (single-turn)\\n\",\n \"import validation_library as veri\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contract_VL_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"b1b01a15\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% SINGLE-TURN: PRE (alpha) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_alpha = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"PRE single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" pre_line, _ = annotations[\\\"pre\\\"]\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"alpha\\\")\\n\",\n \" results_alpha.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, pre_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"pre\\\"]\\n\",\n \"results_alpha += [''] * (len(all_contracts) - len(results_alpha))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_alpha, mapping), flat_entries):\\n\",\n \" line, _ = annotations[\\\"pre\\\"]\\n\",\n \" annotations[\\\"pre\\\"] = (line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contract_PV_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"91ca75eb\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% SINGLE-TURN: POST (omega) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_omega = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"POST single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" post_line, _ = annotations[\\\"post\\\"]\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"omega\\\")\\n\",\n \" results_omega.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, post_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"post\\\"]\\n\",\n \"results_omega += [''] * (len(all_contracts) - len(results_omega))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_omega, mapping), flat_entries):\\n\",\n \" annotations[\\\"post\\\"] = (line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contracts_PVP_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"51f5b999\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# This just shows the three locations we populated for the first item.\\n\",\n \"one_key = next(iter(contracts_with_results.keys()))\\n\",\n \"first_contract, ann = contracts_with_results[one_key][0]\\n\",\n \"ann\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"abfe553d\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": []\n }\n ],\n \"metadata\": {\n \"kernelspec\": {\n \"display_name\": \"Python 3\",\n \"language\": \"python\",\n \"name\": \"python3\"\n },\n \"language_info\": {\n \"codemirror_mode\": {\n \"name\": \"ipython\",\n \"version\": 3\n },\n \"file_extension\": \".py\",\n \"mimetype\": \"text/x-python\",\n \"name\": \"python\",\n \"nbconvert_exporter\": \"python\",\n \"pygments_lexer\": \"ipython3\",\n \"version\": \"3.9.6\"\n }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 5\n}\n"}, "1e1ff481d6bab2cd767f66f16db8e0973afbd683": {"t": 1785753754.053595, "n": 17721, "h": "fcb5e11e0a1ad825", "c": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"86039af7\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"\\n\",\n \"# Requires: pip install openai python-dotenv\\n\",\n \"import os, time, re, json\\n\",\n \"from typing import Optional\\n\",\n \"from dotenv import load_dotenv\\n\",\n \"\\n\",\n \"load_dotenv()\\n\",\n \"\\n\",\n \"try:\\n\",\n \" from openai import OpenAI\\n\",\n \"except Exception as e:\\n\",\n \" raise RuntimeError(\\\"Install the OpenAI SDK first: pip install openai\\\") from e\\n\",\n \"\\n\",\n \"DEEPSEEK_API_KEY = os.environ.get(\\\"DEEPSEEK_API_KEY\\\", \\\"\\\")\\n\",\n \"DEEPSEEK_BASE_URL = os.environ.get(\\\"DEEPSEEK_BASE_URL\\\", \\\"https://api.deepseek.com\\\")\\n\",\n \"DEEPSEEK_MODEL = os.environ.get(\\\"DEEPSEEK_MODEL\\\", \\\"deepseek-chat\\\") # or your preferred\\n\",\n \"\\n\",\n \"if not DEEPSEEK_API_KEY:\\n\",\n \" raise RuntimeError(\\\"Please set DEEPSEEK_API_KEY in your environment (or .env).\\\")\\n\",\n \"\\n\",\n \"client = OpenAI(api_key=DEEPSEEK_API_KEY, base_url=DEEPSEEK_BASE_URL)\\n\",\n \"\\n\",\n \"def _chat_once(prefix: str, suffix: str, temperature=0.8, top_p=0.95, max_tokens=128) -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Sends a single chat completion asking for ONLY the Solidity predicate (no require(...);).\\n\",\n \" Uses a FIM-like instruction by giving prefix/suffix and a <FILL_ME> placeholder.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" prompt = (\\n\",\n \" \\\"You are filling a single-line Solidity predicate to go inside require(...).\\\\n\\\"\\n\",\n \" \\\"Given a source split into PREFIX and SUFFIX around a placeholder `require(<FILL_ME>);`, \\\"\\n\",\n \" \\\"produce ONLY the predicate expression for the require (no semicolon, no backticks, no commentary).\\\\n\\\"\\n\",\n \" \\\"Examples:\\\\n\\\"\\n\",\n \" \\\" - Output like: msg.sender == owner\\\\n\\\"\\n\",\n \" \\\" - NOT like: require(msg.sender == owner);\\\\n\\\"\\n\",\n \" \\\" - NOT like: ```solidity ... ```\\\\n\\\"\\n\",\n \" \\\"\\\\n\\\"\\n\",\n \" \\\"PREFIX:\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" f\\\"{prefix}\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" \\\"SUFFIX:\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" f\\\"{suffix}\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" )\\n\",\n \" resp = client.chat.completions.create(\\n\",\n \" model=DEEPSEEK_MODEL,\\n\",\n \" messages=[\\n\",\n \" {\\\"role\\\": \\\"system\\\", \\\"content\\\": \\\"Return only the Solidity predicate expression.\\\"},\\n\",\n \" {\\\"role\\\": \\\"user\\\", \\\"content\\\": prompt},\\n\",\n \" ],\\n\",\n \" temperature=temperature,\\n\",\n \" top_p=top_p,\\n\",\n \" max_tokens=max_tokens,\\n\",\n \" stop=None,\\n\",\n \" )\\n\",\n \" return (resp.choices[0].message.content or \\\"\\\").strip()\\n\",\n \"\\n\",\n \"def _extract_predicate(text: str) -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Normalize model output to a single-line predicate:\\n\",\n \" - Strip code fences/backticks/semicolons\\n\",\n \" - If it accidentally returned require(...), extract the (...)\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" t = text.strip()\\n\",\n \" t = re.sub(r\\\"^```[a-zA-Z]*\\\\s*\\\", \\\"\\\", t).strip()\\n\",\n \" t = re.sub(r\\\"```$\\\", \\\"\\\", t).strip()\\n\",\n \" # require(...) wrapper?\\n\",\n \" m = re.search(r\\\"require\\\\s*\\\\(\\\\s*(.+?)\\\\s*\\\\)\\\\s*;?\\\\s*$\\\", t, flags=re.S)\\n\",\n \" if m:\\n\",\n \" return m.group(1).strip()\\n\",\n \" # strip trailing semicolon if present\\n\",\n \" t = re.sub(r\\\";+$\\\", \\\"\\\", t).strip()\\n\",\n \" # collapse lines\\n\",\n \" t = \\\" \\\".join(t.splitlines()).strip()\\n\",\n \" return t\\n\",\n \"\\n\",\n \"def deepseek_generate_fim(prefix: str, suffix: str) -> str:\\n\",\n \" # lightweight retry/backoff\\n\",\n \" delays = [0, 1, 2, 4]\\n\",\n \" for d in delays:\\n\",\n \" if d: time.sleep(d)\\n\",\n \" try:\\n\",\n \" raw = _chat_once(prefix, suffix)\\n\",\n \" return _extract_predicate(raw)\\n\",\n \" except Exception:\\n\",\n \" if d == delays[-1]:\\n\",\n \" raise\\n\",\n \" continue\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"8be587ac\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Injection utilities: build masked prefix/suffix at a location (single-turn)\\n\",\n \"import validation_library as veri\\n\",\n \"\\n\",\n \"def _insert_mask_at_line(src: str, line_no_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert `require(<FILL_ME>);` BEFORE 1-based line_no_1based.\\n\",\n \" Returns (prefix, suffix, indent_str).\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" lines = src.splitlines()\\n\",\n \" idx = max(0, min(len(lines), line_no_1based - 1))\\n\",\n \" # Determine indentation: use target line's indentation if exists\\n\",\n \" if 0 <= idx < len(lines):\\n\",\n \" m = re.match(r\\\"^(\\\\s*)\\\", lines[idx])\\n\",\n \" indent = m.group(1) if m else indent_fallback\\n\",\n \" else:\\n\",\n \" indent = indent_fallback\\n\",\n \" masked = f\\\"{indent}require(<FILL_ME>);\\\"\\n\",\n \" prefix = \\\"\\\\n\\\".join(lines[:idx] + [masked])\\n\",\n \" suffix = \\\"\\\\n\\\".join(lines[idx:])\\n\",\n \" return prefix, suffix, indent\\n\",\n \"\\n\",\n \"def _insert_mask_at_function_entry(src: str, vl_line_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert mask as the first statement of the function that contains vl_line_1based.\\n\",\n \" Uses veri.find_function_bounds to resolve the function block.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" pre_line_0based, post_line_0based = veri.find_function_bounds(src, vl_line_1based) # returns 0-based\\n\",\n \" # we want to insert just after the opening brace -> typical pre = opening brace line\\n\",\n \" insert_at = pre_line_0based + 1 # 1-based\\n\",\n \" return _insert_mask_at_line(src, insert_at, indent_fallback=indent_fallback)\\n\",\n \"\\n\",\n \"def _insert_mask_at_function_exit(src: str, vl_line_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert mask just before the closing brace of the function that contains vl_line_1based.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" pre_line_0based, post_line_0based = veri.find_function_bounds(src, vl_line_1based)\\n\",\n \" insert_at = post_line_0based + 1 # convert to 1-based\\n\",\n \" return _insert_mask_at_line(src, insert_at, indent_fallback=indent_fallback)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"b773c049\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Core: single-turn synthesizer for {V, alpha, omega}\\n\",\n \"def synthesize_require_statement(\\n\",\n \" contract_text: str,\\n\",\n \" line_no: int, # VL reference line (1-based)\\n\",\n \" injection: str = \\\"V\\\", # \\\"V\\\" (target line), \\\"alpha\\\" (function entry), \\\"omega\\\" (function exit)\\n\",\n \" indent_fallback: str = \\\" \\\",\\n\",\n \" extra_hint: Optional[str] = None,\\n\",\n \") -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Returns ONLY the predicate (no semicolon). We will wrap as require({pred});\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" if injection == \\\"V\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_line(contract_text, line_no, indent_fallback)\\n\",\n \" elif injection == \\\"alpha\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_function_entry(contract_text, line_no, indent_fallback)\\n\",\n \" elif injection == \\\"omega\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_function_exit(contract_text, line_no, indent_fallback)\\n\",\n \" else:\\n\",\n \" raise ValueError(\\\"injection must be one of: 'V', 'alpha', 'omega'\\\")\\n\",\n \"\\n\",\n \" if extra_hint:\\n\",\n \" prefix = f\\\"// Hint: {extra_hint}\\\\n\\\" + prefix\\n\",\n \"\\n\",\n \" predicate = deepseek_generate_fim(prefix, suffix)\\n\",\n \" return predicate\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"19c25c07\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Seed loader (auto from your aggregated VL artifact)\\n\",\n \"# This replaces the previous seed_contracts-based cell.\\n\",\n \"\\n\",\n \"import os, json\\n\",\n \"from collections import defaultdict\\n\",\n \"\\n\",\n \"# We’ll auto-pick one of these paths unless SEED_VL_JSON is set.\\n\",\n \"CANDIDATES = [\\n\",\n \" os.environ.get(\\\"SEED_VL_JSON\\\", \\\"\\\").strip(),\\n\",\n \" \\\"reports/aggregated/DeepSeek/contract_VL.json\\\",\\n\",\n \" \\\"reports/aggregated/DeepSeek/contracts_VL.json\\\",\\n\",\n \" \\\"reports/aggregated/contract_VL.json\\\",\\n\",\n \"]\\n\",\n \"\\n\",\n \"seed_path = next((p for p in CANDIDATES if p and os.path.exists(p)), None)\\n\",\n \"if not seed_path:\\n\",\n \" raise RuntimeError(\\n\",\n \" \\\"Could not find a VL seed JSON. \\\"\\n\",\n \" \\\"Set SEED_VL_JSON to your file or generate reports/aggregated/DeepSeek/contract_VL.json first.\\\"\\n\",\n \" )\\n\",\n \"\\n\",\n \"with open(seed_path, \\\"r\\\", encoding=\\\"utf-8\\\") as f:\\n\",\n \" raw = json.load(f)\\n\",\n \"\\n\",\n \"# raw is expected to be: { name: [ [contract_text, { 'VL':[line, guard], 'pre':[line, ''], 'post':[line, ''] } ], ... ] }\\n\",\n \"# (Tuples become lists in JSON; guards may be empty strings.)\\n\",\n \"contracts_with_results = defaultdict(list)\\n\",\n \"\\n\",\n \"def _read_line_pair(x):\\n\",\n \" # Accept [line, guard] or just line\\n\",\n \" if isinstance(x, list) and x:\\n\",\n \" return int(x[0]), (x[1] if len(x) > 1 else \\\"\\\")\\n\",\n \" if isinstance(x, int):\\n\",\n \" return int(x), \\\"\\\"\\n\",\n \" return None, \\\"\\\"\\n\",\n \"\\n\",\n \"loaded = 0\\n\",\n \"for name, entries in raw.items():\\n\",\n \" for entry in entries:\\n\",\n \" # entry may be [contract_text, annotations] or a dict with same fields\\n\",\n \" if isinstance(entry, list) and len(entry) == 2:\\n\",\n \" contract_text, ann = entry\\n\",\n \" elif isinstance(entry, dict):\\n\",\n \" contract_text = entry.get(\\\"contract\\\") or entry.get(\\\"contract_text\\\") or entry.get(\\\"0\\\")\\n\",\n \" ann = entry.get(\\\"annotations\\\") or entry.get(\\\"1\\\") or {k: entry.get(k) for k in (\\\"VL\\\",\\\"pre\\\",\\\"post\\\")}\\n\",\n \" else:\\n\",\n \" continue\\n\",\n \"\\n\",\n \" if not isinstance(contract_text, str) or not contract_text.strip():\\n\",\n \" continue\\n\",\n \"\\n\",\n \" vl_line, vl_guard = _read_line_pair((ann or {}).get(\\\"VL\\\"))\\n\",\n \" pre_line, pre_guard = _read_line_pair((ann or {}).get(\\\"pre\\\"))\\n\",\n \" post_line, post_guard = _read_line_pair((ann or {}).get(\\\"post\\\"))\\n\",\n \"\\n\",\n \" # Fallback: recompute pre/post lines if missing using your helper\\n\",\n \" if not isinstance(vl_line, int) or vl_line <= 0:\\n\",\n \" continue\\n\",\n \" try:\\n\",\n \" pre0, post0 = veri.find_function_bounds(contract_text, vl_line) # 0-based\\n\",\n \" if not isinstance(pre_line, int) or pre_line <= 0:\\n\",\n \" pre_line = pre0 + 1 # store function-start (brace) line (0-based+1)\\n\",\n \" # We actually insert at first statement; later logic uses VL+injection to compute exact spot.\\n\",\n \" if not isinstance(post_line, int) or post_line <= 0:\\n\",\n \" post_line = post0 + 1 # closing brace line (0-based+1)\\n\",\n \" except Exception:\\n\",\n \" continue\\n\",\n \"\\n\",\n \" contracts_with_results[name].append((\\n\",\n \" contract_text,\\n\",\n \" {\\n\",\n \" \\\"VL\\\": (vl_line, vl_guard or \\\"\\\"),\\n\",\n \" \\\"pre\\\": (pre_line, pre_guard or \\\"\\\"),\\n\",\n \" \\\"post\\\":(post_line, post_guard or \\\"\\\"),\\n\",\n \" }\\n\",\n \" ))\\n\",\n \" loaded += 1\\n\",\n \"\\n\",\n \"print(f\\\"[seed] Loaded {loaded} contract entries from {seed_path}\\\")\\n\",\n \"len(contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"a8d85734\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Output directory for SINGLE-TURN results\\n\",\n \"os.makedirs(\\\"reports/isolated/DeepSeek\\\", exist_ok=True)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"8461a3bb\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% SINGLE-TURN: Vulnerable-line (V) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_V = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"VL single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" # SINGLE-TURN: use ORIGINAL contract (no prior injections)\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"V\\\")\\n\",\n \" results_V.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, vl_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"VL\\\"]\\n\",\n \"results_V += [''] * (len(all_contracts) - len(results_V))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, vl_line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_V, mapping), flat_entries):\\n\",\n \" annotations[\\\"VL\\\"] = (vl_line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"# Minimal VL-only drop (single-turn)\\n\",\n \"import validation_library as veri\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contract_VL_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"b1b01a15\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% SINGLE-TURN: PRE (alpha) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_alpha = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"PRE single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" pre_line, _ = annotations[\\\"pre\\\"]\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"alpha\\\")\\n\",\n \" results_alpha.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, pre_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"pre\\\"]\\n\",\n \"results_alpha += [''] * (len(all_contracts) - len(results_alpha))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_alpha, mapping), flat_entries):\\n\",\n \" line, _ = annotations[\\\"pre\\\"]\\n\",\n \" annotations[\\\"pre\\\"] = (line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contract_PV_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"91ca75eb\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% SINGLE-TURN: POST (omega) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_omega = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"POST single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" post_line, _ = annotations[\\\"post\\\"]\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"omega\\\")\\n\",\n \" results_omega.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, post_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"post\\\"]\\n\",\n \"results_omega += [''] * (len(all_contracts) - len(results_omega))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_omega, mapping), flat_entries):\\n\",\n \" annotations[\\\"post\\\"] = (line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contracts_PVP_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"51f5b999\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# This just shows the three locations we populated for the first item.\\n\",\n \"one_key = next(iter(contracts_with_results.keys()))\\n\",\n \"first_contract, ann = contracts_with_results[one_key][0]\\n\",\n \"ann\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"abfe553d\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": []\n }\n ],\n \"metadata\": {\n \"kernelspec\": {\n \"display_name\": \"Python 3\",\n \"language\": \"python\",\n \"name\": \"python3\"\n },\n \"language_info\": {\n \"codemirror_mode\": {\n \"name\": \"ipython\",\n \"version\": 3\n },\n \"file_extension\": \".py\",\n \"mimetype\": \"text/x-python\",\n \"name\": \"python\",\n \"nbconvert_exporter\": \"python\",\n \"pygments_lexer\": \"ipython3\",\n \"version\": \"3.9.6\"\n }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 5\n}\n"}, "a317e12fe6f706ae86f1043e9e812657cfa3c0d9": {"t": 1785753755.5528023, "n": 19024, "h": "a110caf50ddec4b7", "c": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": 4,\n \"id\": \"86039af7\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"\\n\",\n \"# Requires: pip install openai python-dotenv\\n\",\n \"import os, time, re, json\\n\",\n \"from typing import Optional\\n\",\n \"from dotenv import load_dotenv\\n\",\n \"\\n\",\n \"load_dotenv()\\n\",\n \"\\n\",\n \"try:\\n\",\n \" from openai import OpenAI\\n\",\n \"except Exception as e:\\n\",\n \" raise RuntimeError(\\\"Install the OpenAI SDK first: pip install openai\\\") from e\\n\",\n \"\\n\",\n \"DEEPSEEK_API_KEY = os.environ.get(\\\"DEEPSEEK_API_KEY\\\", \\\"\\\")\\n\",\n \"DEEPSEEK_BASE_URL = os.environ.get(\\\"DEEPSEEK_BASE_URL\\\", \\\"https://api.deepseek.com\\\")\\n\",\n \"DEEPSEEK_MODEL = os.environ.get(\\\"DEEPSEEK_MODEL\\\", \\\"deepseek-chat\\\") # or your preferred\\n\",\n \"\\n\",\n \"if not DEEPSEEK_API_KEY:\\n\",\n \" raise RuntimeError(\\\"Please set DEEPSEEK_API_KEY in your environment (or .env).\\\")\\n\",\n \"\\n\",\n \"client = OpenAI(api_key=DEEPSEEK_API_KEY, base_url=DEEPSEEK_BASE_URL)\\n\",\n \"\\n\",\n \"def _chat_once(prefix: str, suffix: str, temperature=0.8, top_p=0.95, max_tokens=128) -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Sends a single chat completion asking for ONLY the Solidity predicate (no require(...);).\\n\",\n \" Uses a FIM-like instruction by giving prefix/suffix and a <FILL_ME> placeholder.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" prompt = (\\n\",\n \" \\\"You are filling a single-line Solidity predicate to go inside require(...).\\\\n\\\"\\n\",\n \" \\\"Given a source split into PREFIX and SUFFIX around a placeholder `require(<FILL_ME>);`, \\\"\\n\",\n \" \\\"produce ONLY the predicate expression for the require (no semicolon, no backticks, no commentary).\\\\n\\\"\\n\",\n \" \\\"Examples:\\\\n\\\"\\n\",\n \" \\\" - Output like: msg.sender == owner\\\\n\\\"\\n\",\n \" \\\" - NOT like: require(msg.sender == owner);\\\\n\\\"\\n\",\n \" \\\" - NOT like: ```solidity ... ```\\\\n\\\"\\n\",\n \" \\\"\\\\n\\\"\\n\",\n \" \\\"PREFIX:\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" f\\\"{prefix}\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" \\\"SUFFIX:\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" f\\\"{suffix}\\\\n\\\"\\n\",\n \" \\\"-----\\\\n\\\"\\n\",\n \" )\\n\",\n \" resp = client.chat.completions.create(\\n\",\n \" model=DEEPSEEK_MODEL,\\n\",\n \" messages=[\\n\",\n \" {\\\"role\\\": \\\"system\\\", \\\"content\\\": \\\"Return only the Solidity predicate expression.\\\"},\\n\",\n \" {\\\"role\\\": \\\"user\\\", \\\"content\\\": prompt},\\n\",\n \" ],\\n\",\n \" temperature=temperature,\\n\",\n \" top_p=top_p,\\n\",\n \" max_tokens=max_tokens,\\n\",\n \" stop=None,\\n\",\n \" )\\n\",\n \" return (resp.choices[0].message.content or \\\"\\\").strip()\\n\",\n \"\\n\",\n \"def _extract_predicate(text: str) -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Normalize model output to a single-line predicate:\\n\",\n \" - Strip code fences/backticks/semicolons\\n\",\n \" - If it accidentally returned require(...), extract the (...)\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" t = text.strip()\\n\",\n \" t = re.sub(r\\\"^```[a-zA-Z]*\\\\s*\\\", \\\"\\\", t).strip()\\n\",\n \" t = re.sub(r\\\"```$\\\", \\\"\\\", t).strip()\\n\",\n \" # require(...) wrapper?\\n\",\n \" m = re.search(r\\\"require\\\\s*\\\\(\\\\s*(.+?)\\\\s*\\\\)\\\\s*;?\\\\s*$\\\", t, flags=re.S)\\n\",\n \" if m:\\n\",\n \" return m.group(1).strip()\\n\",\n \" # strip trailing semicolon if present\\n\",\n \" t = re.sub(r\\\";+$\\\", \\\"\\\", t).strip()\\n\",\n \" # collapse lines\\n\",\n \" t = \\\" \\\".join(t.splitlines()).strip()\\n\",\n \" return t\\n\",\n \"\\n\",\n \"def deepseek_generate_fim(prefix: str, suffix: str) -> str:\\n\",\n \" # lightweight retry/backoff\\n\",\n \" delays = [0, 1, 2, 4]\\n\",\n \" for d in delays:\\n\",\n \" if d: time.sleep(d)\\n\",\n \" try:\\n\",\n \" raw = _chat_once(prefix, suffix)\\n\",\n \" return _extract_predicate(raw)\\n\",\n \" except Exception:\\n\",\n \" if d == delays[-1]:\\n\",\n \" raise\\n\",\n \" continue\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": 5,\n \"id\": \"8be587ac\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Injection utilities: build masked prefix/suffix at a location (single-turn)\\n\",\n \"import validation_library as veri\\n\",\n \"\\n\",\n \"def _insert_mask_at_line(src: str, line_no_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert `require(<FILL_ME>);` BEFORE 1-based line_no_1based.\\n\",\n \" Returns (prefix, suffix, indent_str).\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" lines = src.splitlines()\\n\",\n \" idx = max(0, min(len(lines), line_no_1based - 1))\\n\",\n \" # Determine indentation: use target line's indentation if exists\\n\",\n \" if 0 <= idx < len(lines):\\n\",\n \" m = re.match(r\\\"^(\\\\s*)\\\", lines[idx])\\n\",\n \" indent = m.group(1) if m else indent_fallback\\n\",\n \" else:\\n\",\n \" indent = indent_fallback\\n\",\n \" masked = f\\\"{indent}require(<FILL_ME>);\\\"\\n\",\n \" prefix = \\\"\\\\n\\\".join(lines[:idx] + [masked])\\n\",\n \" suffix = \\\"\\\\n\\\".join(lines[idx:])\\n\",\n \" return prefix, suffix, indent\\n\",\n \"\\n\",\n \"def _insert_mask_at_function_entry(src: str, vl_line_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert mask as the first statement of the function that contains vl_line_1based.\\n\",\n \" Uses veri.find_function_bounds to resolve the function block.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" pre_line_0based, post_line_0based = veri.find_function_bounds(src, vl_line_1based) # returns 0-based\\n\",\n \" # we want to insert just after the opening brace -> typical pre = opening brace line\\n\",\n \" insert_at = pre_line_0based + 1 # 1-based\\n\",\n \" return _insert_mask_at_line(src, insert_at, indent_fallback=indent_fallback)\\n\",\n \"\\n\",\n \"def _insert_mask_at_function_exit(src: str, vl_line_1based: int, indent_fallback=\\\" \\\"):\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Insert mask just before the closing brace of the function that contains vl_line_1based.\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" pre_line_0based, post_line_0based = veri.find_function_bounds(src, vl_line_1based)\\n\",\n \" insert_at = post_line_0based + 1 # convert to 1-based\\n\",\n \" return _insert_mask_at_line(src, insert_at, indent_fallback=indent_fallback)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": 6,\n \"id\": \"b773c049\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Core: single-turn synthesizer for {V, alpha, omega}\\n\",\n \"def synthesize_require_statement(\\n\",\n \" contract_text: str,\\n\",\n \" line_no: int, # VL reference line (1-based)\\n\",\n \" injection: str = \\\"V\\\", # \\\"V\\\" (target line), \\\"alpha\\\" (function entry), \\\"omega\\\" (function exit)\\n\",\n \" indent_fallback: str = \\\" \\\",\\n\",\n \" extra_hint: Optional[str] = None,\\n\",\n \") -> str:\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" Returns ONLY the predicate (no semicolon). We will wrap as require({pred});\\n\",\n \" \\\"\\\"\\\"\\n\",\n \" if injection == \\\"V\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_line(contract_text, line_no, indent_fallback)\\n\",\n \" elif injection == \\\"alpha\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_function_entry(contract_text, line_no, indent_fallback)\\n\",\n \" elif injection == \\\"omega\\\":\\n\",\n \" prefix, suffix, _ = _insert_mask_at_function_exit(contract_text, line_no, indent_fallback)\\n\",\n \" else:\\n\",\n \" raise ValueError(\\\"injection must be one of: 'V', 'alpha', 'omega'\\\")\\n\",\n \"\\n\",\n \" if extra_hint:\\n\",\n \" prefix = f\\\"// Hint: {extra_hint}\\\\n\\\" + prefix\\n\",\n \"\\n\",\n \" predicate = deepseek_generate_fim(prefix, suffix)\\n\",\n \" return predicate\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": 10,\n \"id\": \"19c25c07\",\n \"metadata\": {},\n \"outputs\": [\n {\n \"name\": \"stdout\",\n \"output_type\": \"stream\",\n \"text\": [\n \"[seed] Loaded 170 contract entries from reports/aggregated/DeepSeek/contract_VL.json\\n\"\n ]\n },\n {\n \"data\": {\n \"text/plain\": [\n \"130\"\n ]\n },\n \"execution_count\": 10,\n \"metadata\": {},\n \"output_type\": \"execute_result\"\n }\n ],\n \"source\": [\n \"# %% Seed loader (auto from your aggregated VL artifact)\\n\",\n \"# This replaces the previous seed_contracts-based cell.\\n\",\n \"\\n\",\n \"import os, json\\n\",\n \"from collections import defaultdict\\n\",\n \"\\n\",\n \"# We’ll auto-pick one of these paths unless SEED_VL_JSON is set.\\n\",\n \"CANDIDATES = [\\n\",\n \" os.environ.get(\\\"SEED_VL_JSON\\\", \\\"\\\").strip(),\\n\",\n \" \\\"reports/aggregated/DeepSeek/contract_VL.json\\\",\\n\",\n \" \\\"reports/aggregated/DeepSeek/contracts_VL.json\\\",\\n\",\n \" \\\"reports/aggregated/contract_VL.json\\\",\\n\",\n \"]\\n\",\n \"\\n\",\n \"seed_path = next((p for p in CANDIDATES if p and os.path.exists(p)), None)\\n\",\n \"if not seed_path:\\n\",\n \" raise RuntimeError(\\n\",\n \" \\\"Could not find a VL seed JSON. \\\"\\n\",\n \" \\\"Set SEED_VL_JSON to your file or generate reports/aggregated/DeepSeek/contract_VL.json first.\\\"\\n\",\n \" )\\n\",\n \"\\n\",\n \"with open(seed_path, \\\"r\\\", encoding=\\\"utf-8\\\") as f:\\n\",\n \" raw = json.load(f)\\n\",\n \"\\n\",\n \"# raw is expected to be: { name: [ [contract_text, { 'VL':[line, guard], 'pre':[line, ''], 'post':[line, ''] } ], ... ] }\\n\",\n \"# (Tuples become lists in JSON; guards may be empty strings.)\\n\",\n \"contracts_with_results = defaultdict(list)\\n\",\n \"\\n\",\n \"def _read_line_pair(x):\\n\",\n \" # Accept [line, guard] or just line\\n\",\n \" if isinstance(x, list) and x:\\n\",\n \" return int(x[0]), (x[1] if len(x) > 1 else \\\"\\\")\\n\",\n \" if isinstance(x, int):\\n\",\n \" return int(x), \\\"\\\"\\n\",\n \" return None, \\\"\\\"\\n\",\n \"\\n\",\n \"loaded = 0\\n\",\n \"for name, entries in raw.items():\\n\",\n \" for entry in entries:\\n\",\n \" # entry may be [contract_text, annotations] or a dict with same fields\\n\",\n \" if isinstance(entry, list) and len(entry) == 2:\\n\",\n \" contract_text, ann = entry\\n\",\n \" elif isinstance(entry, dict):\\n\",\n \" contract_text = entry.get(\\\"contract\\\") or entry.get(\\\"contract_text\\\") or entry.get(\\\"0\\\")\\n\",\n \" ann = entry.get(\\\"annotations\\\") or entry.get(\\\"1\\\") or {k: entry.get(k) for k in (\\\"VL\\\",\\\"pre\\\",\\\"post\\\")}\\n\",\n \" else:\\n\",\n \" continue\\n\",\n \"\\n\",\n \" if not isinstance(contract_text, str) or not contract_text.strip():\\n\",\n \" continue\\n\",\n \"\\n\",\n \" vl_line, vl_guard = _read_line_pair((ann or {}).get(\\\"VL\\\"))\\n\",\n \" pre_line, pre_guard = _read_line_pair((ann or {}).get(\\\"pre\\\"))\\n\",\n \" post_line, post_guard = _read_line_pair((ann or {}).get(\\\"post\\\"))\\n\",\n \"\\n\",\n \" # Fallback: recompute pre/post lines if missing using your helper\\n\",\n \" if not isinstance(vl_line, int) or vl_line <= 0:\\n\",\n \" continue\\n\",\n \" try:\\n\",\n \" pre0, post0 = veri.find_function_bounds(contract_text, vl_line) # 0-based\\n\",\n \" if not isinstance(pre_line, int) or pre_line <= 0:\\n\",\n \" pre_line = pre0 + 1 # store function-start (brace) line (0-based+1)\\n\",\n \" # We actually insert at first statement; later logic uses VL+injection to compute exact spot.\\n\",\n \" if not isinstance(post_line, int) or post_line <= 0:\\n\",\n \" post_line = post0 + 1 # closing brace line (0-based+1)\\n\",\n \" except Exception:\\n\",\n \" continue\\n\",\n \"\\n\",\n \" contracts_with_results[name].append((\\n\",\n \" contract_text,\\n\",\n \" {\\n\",\n \" \\\"VL\\\": (vl_line, vl_guard or \\\"\\\"),\\n\",\n \" \\\"pre\\\": (pre_line, pre_guard or \\\"\\\"),\\n\",\n \" \\\"post\\\":(post_line, post_guard or \\\"\\\"),\\n\",\n \" }\\n\",\n \" ))\\n\",\n \" loaded += 1\\n\",\n \"\\n\",\n \"print(f\\\"[seed] Loaded {loaded} contract entries from {seed_path}\\\")\\n\",\n \"len(contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": 11,\n \"id\": \"a8d85734\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n \"# %% Output directory for SINGLE-TURN results\\n\",\n \"os.makedirs(\\\"reports/isolated/DeepSeek\\\", exist_ok=True)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": 12,\n \"id\": \"8461a3bb\",\n \"metadata\": {},\n \"outputs\": [\n {\n \"name\": \"stderr\",\n \"output_type\": \"stream\",\n \"text\": [\n \"VL single-turn: 0%| | 0/130 [00:00<?, ?it/s]\"\n ]\n },\n {\n \"name\": \"stderr\",\n \"output_type\": \"stream\",\n \"text\": [\n \"VL single-turn: 100%|██████████| 130/130 [10:31<00:00, 4.86s/it]\\n\"\n ]\n }\n ],\n \"source\": [\n \"# %% SINGLE-TURN: Vulnerable-line (V) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_V = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"VL single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" # SINGLE-TURN: use ORIGINAL contract (no prior injections)\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"V\\\")\\n\",\n \" results_V.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, vl_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"VL\\\"]\\n\",\n \"results_V += [''] * (len(all_contracts) - len(results_V))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, vl_line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_V, mapping), flat_entries):\\n\",\n \" annotations[\\\"VL\\\"] = (vl_line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"# Minimal VL-only drop (single-turn)\\n\",\n \"import validation_library as veri\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contract_VL_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": 13,\n \"id\": \"b1b01a15\",\n \"metadata\": {},\n \"outputs\": [\n {\n \"name\": \"stderr\",\n \"output_type\": \"stream\",\n \"text\": [\n \"PRE single-turn: 100%|██████████| 130/130 [10:24<00:00, 4.81s/it]\\n\"\n ]\n }\n ],\n \"source\": [\n \"# %% SINGLE-TURN: PRE (alpha) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_alpha = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"PRE single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" pre_line, _ = annotations[\\\"pre\\\"]\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"alpha\\\")\\n\",\n \" results_alpha.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, pre_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"pre\\\"]\\n\",\n \"results_alpha += [''] * (len(all_contracts) - len(results_alpha))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_alpha, mapping), flat_entries):\\n\",\n \" line, _ = annotations[\\\"pre\\\"]\\n\",\n \" annotations[\\\"pre\\\"] = (line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contract_PV_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": 14,\n \"id\": \"91ca75eb\",\n \"metadata\": {},\n \"outputs\": [\n {\n \"name\": \"stderr\",\n \"output_type\": \"stream\",\n \"text\": [\n \"POST single-turn: 100%|██████████| 130/130 [10:23<00:00, 4.79s/it]\\n\"\n ]\n }\n ],\n \"source\": [\n \"# %% SINGLE-TURN: POST (omega) inference (original contract only)\\n\",\n \"from tqdm import tqdm\\n\",\n \"\\n\",\n \"all_contracts = []\\n\",\n \"mapping = []\\n\",\n \"results_omega = []\\n\",\n \"\\n\",\n \"for contract_name, entries in tqdm(contracts_with_results.items(), desc=\\\"POST single-turn\\\"):\\n\",\n \" for contract, annotations in entries:\\n\",\n \" vl_line, _ = annotations[\\\"VL\\\"]\\n\",\n \" post_line, _ = annotations[\\\"post\\\"]\\n\",\n \" pred = synthesize_require_statement(contract, line_no=vl_line, injection=\\\"omega\\\")\\n\",\n \" results_omega.append(pred)\\n\",\n \" all_contracts.append(contract)\\n\",\n \" mapping.append((contract_name, contract, post_line))\\n\",\n \"\\n\",\n \"# write back to annotations[\\\"post\\\"]\\n\",\n \"results_omega += [''] * (len(all_contracts) - len(results_omega))\\n\",\n \"flat_entries = [(cn, entry) for cn, es in contracts_with_results.items() for entry in es]\\n\",\n \"\\n\",\n \"for (generated, (contract_name, contract, line)), (_, (existing_contract, annotations)) in \\\\\\n\",\n \" zip(zip(results_omega, mapping), flat_entries):\\n\",\n \" annotations[\\\"post\\\"] = (line, f\\\"require({generated});\\\")\\n\",\n \"\\n\",\n \"veri.print_json_report(\\\"reports/isolated/DeepSeek/contracts_PVP_single.json\\\", contracts_with_results)\\n\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"51f5b999\",\n \"metadata\": {},\n \"outputs\": [\n {\n \"data\": {\n \"text/plain\": [\n \"{'VL': (11, 'require(max > 0);'),\\n\",\n \" 'pre': (11, 'require(salt == 0);'),\\n\",\n \" 'post': (26, 'require(msg.sender == address(this));')}\"\n ]\n },\n \"execution_count\": 15,\n \"metadata\": {},\n \"output_type\": \"execute_result\"\n }\n ],\n \"source\": [\n \"# This just shows the three locations we populated for the first item.\\n\",\n \"one_key = next(iter(contracts_with_results.keys()))\\n\",\n \"first_contract, ann = contracts_with_results[one_key][0]\\n\",\n \"ann\"\n ]\n },\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"id\": \"abfe553d\",\n \"metadata\": {},\n \"outputs\": [],\n \"source\": []\n }\n ],\n \"metadata\": {\n \"kernelspec\": {\n \"display_name\": \"Python 3\",\n \"language\": \"python\",\n \"name\": \"python3\"\n },\n \"language_info\": {\n \"codemirror_mode\": {\n \"name\": \"ipython\",\n \"version\": 3\n },\n \"file_extension\": \".py\",\n \"mimetype\": \"text/x-python\",\n \"name\": \"python\",\n \"nbconvert_exporter\": \"python\",\n \"pygments_lexer\": \"ipython3\",\n \"version\": \"3.9.6\"\n }\n },\n \"nbformat\": 4,\n \"nbformat_minor\": 5\n}\n"}}}} |